Edward Glaeser on Joblessness and the War on Work
What this covers
Ed Glaeser of Harvard University and Russ Roberts explore why employment among prime-age men has collapsed over recent decades and whether a Universal Basic Income offers a solution. Glaeser argues that the crisis stems from a two-sided squeeze: collapsing labor demand for less-educated workers as manufacturing declined and innovation shifted toward skilled labor, combined with reduced labor supply as disability insurance and other benefit programs lowered the incentive to work. The conversation pivots on a crucial distinction: joblessness—permanent exit from the labor force—differs sharply from unemployment, and the data show most jobless men are not idle by necessity but by choice, yet their time away from work correlates strongly with opioid use, divorce, and despair rather than with home care or community engagement.
Glaeser moves through several concrete policy areas. He documents how geographic mobility has collapsed since 2007 and how restrictive housing markets in prosperous regions prevent less-skilled workers from accessing higher-wage areas, reversing the postwar convergence that lifted poor regions. He argues that modern innovation, unlike Henry Ford's model, creates complementary demand for high-skilled workers rather than mass low-skill employment, and points to occupational licensing and regulatory barriers as underappreciated suppressors of poor-person entrepreneurship. On education, he resists the blanket case for more college attendance, instead advocating competitively sourced vocational training. Most contentiously, Glaeser positions UBI as a failure of imagination: paying 30 to 40 percent of Americans not to work would breed resentment and despair because work supplies purpose that cash cannot, a claim grounded in time-use data and Second-Order Thinking about what humans actually value beyond material consumption.
Glaeser argues that the rise of joblessness among prime-age men is America's central domestic crisis driven by both falling labor demand for the less-educated and a reduced labor supply (more generous benefits and reduced mobility), and that a Universal Basic Income would worsen rather than solve it because work provides meaning that cash transfers cannot.
- Joblessness, not low wages, is most strongly associated with misery, opioid use, and despair
- Modern innovation complements skilled workers rather than creating mass low-skill employment, while housing restrictions and falling migration trap workers in distressed regions
- Work supplies purpose and social connection that a check cannot replace, so UBI risks creating a resentful, idle 30-40%
Joblessness among less-educated prime-age men has become severe and is widening sharply by education and region.
- Joblessness is distinct from unemployment: the unemployed lack a job but are actively looking, whereas most of the jobless (more than two-thirds) have left the labor force entirely and say they are no longer interested in employment, though economists assume everyone would take a job at a sufficiently high wage.
“unemployment means that you lack a job but are actively looking for work... less than a third of those people what would qualify as jobless. The remainder are people who have left the labor force entirely.”
- The rise of joblessness, especially among prime-age men, is the great American domestic crisis of the twenty-first century because the lives of the long-term jobless are extraordinarily tough, marked disproportionately by lower life satisfaction, higher divorce, opioid use, and disability, and because work provides a sense of purpose that money alone cannot.
“The rise of joblessness--especially among men--is the great American domestic crisis of the twenty-first century.”
- In Glaeser's data with Austin and Summers (pooled 2010-2015 ASEC), 18.6 percent of prime-age men were jobless, but only 2.8 percentage points lived alone; the largest group (10.1 percent of all men) were long-term jobless living with others, with average family income around $42,000 of which they contributed only about $8,000 (over half from disability), meaning most jobless men survive on the income of a spouse or, surprisingly often, parents rather than living in dire isolation.
“we have 18.6% of men who are jobless in our sample. Of this, only 2.8 percentage points of the total population are people who are living alone. The other 15.8% are living with somebody else.”
- Joblessness is disproportionately a condition of the poorly educated, and the employment-rate gap between the most and least educated prime-age groups has widened from about 6 percentage points in 1977 to almost 15 percentage points today, with enormous regional variation (Kentucky leading at roughly 23 percent male jobless versus under 10 percent in Iowa).
“The employment-rate gap between the most and least educated groups has widened from about 6 percent in 1977 to almost 15 percent today.”
- American geographic mobility has collapsed: from 1950 to 1992 the inter-county migration rate never fell below 6 percent annually, but since 2007 it has never risen above 4 percent, and the nature of migration also changed so that, unlike pre-1960 when people moved toward higher-income areas, less-skilled Americans now rarely migrate toward high-income regions.
“between 1950 and 1992, the inter-county migration rate... was never less than 6%... Over the past 10 years, since 2007, it has never been above 4%.”
Work provides purpose, social connection, and identity that money alone cannot replace, regardless of income level.
- A Universal Basic Income supporting 30-40 percent of America in joblessness is a horror story rather than a solution, not mainly because the working 60-70 percent must pay for it, but because telling that 30 percent their lives will not be ones of contribution or producing anything anyone values would breed boredom, resentment, and hostility toward the check-writers, reflecting a profound misunderstanding of what humans value, since happiness does not come from merely receiving a check.
“it's a horror for the 30%. It's not just that the 70% have to pay for it... How does that 30% not become, does not, bored, become incredibly resentful and hostile towards the 70% that write the checks?”
- The data contradict the idea that giving the jobless enough money makes them content: jobless men living with others have family incomes around $42,000 yet report low life satisfaction and high suicide rates, suggesting the Huxley 'soma' vision of contented idleness is not sensible and that work matters partly because it surrounds people with mates and social connection and positive feedback.
“their family income is $42,000... and yet, their self-reported life satisfaction is pretty low. Their suicide rates are pretty high. It's hard to look at the data and be convinced that the Huxley vision was particularly sensible.”
- Time-use data show that jobless men do not spend their freed-up time on home care or socializing the way jobless women do; in the Eastern Heartland, the jobless have roughly 360 extra minutes a day from not working but allocate only about 20 minutes to socializing while spending around 160 extra minutes watching television, and divorce rates have spiked for this population, undermining the notion that men withdraw from work to contribute at home.
“the overwhelming time use is in watching television, in terms of the big shift. Not in terms of taking care of the kids.”
- Opioid use and mortality are more strongly tied to joblessness than to low wages or income inequality, because someone working a low-wage 40-hour job knows opioid use risks the job, whereas the genuinely jobless are 'looking to fill up an ocean of time and hopelessness'; the share of a county's population on disability in 1990 is a strong predictor of subsequent opioid use and mortality, with the Eastern Heartland as Ground Zero.
“the people who are on the economic margins would seem to me to be far less people at risk than the people who are genuinely jobless, who are just looking to fill up an ocean of time and hopelessness.”
Joblessness results from both collapsing demand for less-educated labor and policy-driven reductions in work incentives.
- Unlike early-20th-century innovators such as Henry Ford who created hundreds of thousands of low-skill jobs, today's great innovators like Bill Gates create products that disproportionately employ high-skilled software engineers, so modern innovation produces complementarity between skilled workers and other skilled workers rather than between skilled and unskilled workers, leaving each non-employed American as a 'failure of entrepreneurial imagination' because the returns to innovating for the skilled are so much higher.
“whereas Henry Ford innovated in a way that provided tens, hundreds of thousands of jobs for less-skilled Americans... Bill Gates innovates in a way that employs disproportionately high-skilled software engineers.”
- Prime-age male joblessness has two primary causes operating together: collapsing labor demand for less-educated men (as factory and mining work declined and innovation favored skilled labor) combined with reduced labor supply, because benefit programs like disability insurance and reliance on working spouses or parents lowered the cost of being jobless while public benefits tied to earnings raised the implicit cost of working.
“the two primary potential culprits are labor demand and labor supply.”
- Social welfare programs such as Disability, Section 8 housing vouchers, and SNAP impose high implicit taxes on earnings that discourage work, so policy should redesign them while simultaneously ramping up employment subsidies given to the worker or the firm in a way that obviously raises the effective wage, like an 'earned income tax credit on steroids.'
“Section 8 Housing Vouchers, SNAP... Food Stamp assistance--all come with implicit taxes of earnings that can easily get quite high.”
- Being formally on Disability does not mean one literally cannot work, but earning above a low threshold causes loss of disability payments, producing an enormous implicit tax (effectively more than 100 percent) on working, which discourages disability recipients from re-entering employment.
“you lose your disability payments once you get over the threshold. So, there's a very mild threshold. Once you get over it, so the tax is enormous. The implicit tax on Disability is enormous.”
Migration barriers and social isolation trap distressed workers in low-opportunity regions rather than enabling geographic mobility.
- Postwar America saw remarkable income convergence as poorer regions caught up: in 1950 there were 18 states with double Mississippi's per-capita income, but today no state has double Mississippi's income because it caught up through both out-migration of workers and in-migration of business and capital; this convergence has now largely stalled because there is much less interest in moving people or capital to low-cost areas, and migrating capital often no longer employs many workers.
“In 1950, there were 18 other states that had double the income of Mississippi. Today, there's not a single state in the Union with double, a per-capita income that's double that in Mississippi. Because it has caught up.”
- Migration is heavily driven by social networks (chain migration), where one migrant to an area convinces people from home to follow; the evidence for this is overwhelming, illustrated by the African American Great Migration north, Appalachian communities clustering in the same urban areas, and retirees moving to be near friends, which means distressed communities lacking ties to prosperous places are unlikely to migrate there.
“the evidence for chain migration, which is one version of what you are saying where, one migrant to an area then talks to his people at home and convinces them to come. The evidence for that is overwhelming.”
- For the first time in American history we have made it hard to move into the most economically successful regions, because restrictive housing markets in places like Silicon Valley make the overall deal unattractive for less-skilled workers; unlike 19th-century settlers or 1900-era tenement immigrants who faced no such barriers, today's restrictive housing policy is a genuine policy problem suppressing beneficial migration.
“Now, you come to Silicon Valley and you face one of the most restrictive housing markets in the country. And I think that's a policy problem.”
- There is a real urban wage premium that reflects cities making workers more productive rather than merely selecting smarter people: studies of migrants (notably De la Roca and Puga's Spanish administrative data) show that people who move to cities experience faster wage growth year by year and carry that gain with them when they leave, and randomized refugee-placement experiments in Sweden also show significant place effects, indicating cities are places where people 'get smart by being around other smart people.'
“people who come to cities experience faster wage growth when they are there... once you show up in the city, year by year, month by month, you experience faster wage growth. And then you typically take that away with you if you move to another city.”
- The urban wage premium must be split into two concepts: the nominal premium (people in cities earn more regular dollars, interesting because it implies higher marginal productivity) and the real premium (dollars corrected for local cost of living); only the real premium is relevant to understanding migration decisions, because a higher nominal wage eaten by higher housing costs gives no reason to move.
“one way of thinking about the urban wage premium is just the straight, nominal fact that people in cities earn more regular dollars. The second of which is the real urban wage premium, meaning the dollars corrected for the local cost of living.”
Vocational training, regulatory deregulation, and employment subsidies should replace college expansion and unconditional income support.
- To improve education specifically for the jobless, the focus should be on vocational training delivered through competitively sourced, possibly non-traditional providers teaching trades like plumbing and carpentry, with programs designed so their effectiveness can actually be evaluated at graduation, given that we don't yet know what works.
“I would be focused on vocational training. And, you know, starting on the fact that we don't know exactly what works, setting up programs where you actually are going to be able to evaluate whether or not the program worked upon the point of graduation.”
- It is outrageous that the United States regulates the entrepreneurship of poor people far more strictly than that of the rich: an internet startup in Cambridge or Palo Alto can reach 50 million users with almost no regulatory oversight, while opening a small grocery selling milk products requires walking through 17 permits in many large cities, so reducing such regulatory barriers, including occupational licensing and adopting one-stop permitting, should be a central pro-employment policy.
“it's outrageous in this country that we regulate the entrepreneurship of poor people so much more strictly than we regulate the entrepreneurship of rich people.”
- Simply sending more people to college will not solve joblessness; education is part of the solution only if it provides skills actually valued in the labor force, which may require better community colleges, alternatives to them, and competitively sourced vocational training, while the simplistic view that raising the college-going rate a few percentage points fixes everything is wrong.
“education needs to do a much better job of providing skills that are actually valued in the labor force. So, if the view is we're just going to send everyone off to college and let them major in concentrations that have absolutely no value in the workforce, I don't see how that helps anything.”
- When cities and communities bid to attract businesses (such as Amazon warehouses or headquarters), they should evaluate not how many jobs are created in the short run but whether the investment will generate new business formation in the long run, since a warehouse may deliver only a few jobs unless it spins off related businesses.
“when cities and communities bid on businesses, they really need to ask themselves not how many jobs are being created in the short run, but whether or not this is going to create new business formation in the long run.”
Service-economy jobs require interpersonal skills that some long-term jobless workers lack, creating a skills mismatch separate from education.
- As AI and automation advance, the occupations most resistant to displacement will be those combining cognitive ability with interpersonal skills, because tasks that merely involve being smart are easier to outsource to a computer, whereas making people feel comfortable and bringing joy to fellow humans remains robust; David Deming's work shows that over the last 15 years even within skill-intensive occupations, the ones that did well required social interaction, not just intelligence.
“David Deming's work has shown that if you look at wage growth over the last 15 years, even in the STEM-intensive, the skill-intensive occupations, the occupations that have done well are those that actually involve interpersonal interactions as well as being smart.”
- A specific obstacle for prime-age male joblessness is the 'curmudgeon effect': a 50-year-old man could get away with not being nice to others while working a factory, mining, or tenured academic job for decades, but the same lack of social skill is a serious handicap in the service economy, so retraining and schooling alone cannot fully solve the problem.
“One of the problems that 50-year-old men often have is they're not great at being nice to other people.”
- The convention that 'prime age' for labor force purposes ends at 55 is offensive.
“I think we want to be on record as saying that prime age stops at 55 is offensive.”
Occupational licensing correlates with higher joblessness and represents excessive regulation of poor people's entrepreneurship.
- Across regions, places with a more heavily licensed workforce are associated with higher joblessness in Glaeser's Brookings regressions, a relationship significant across the past four decades, though he cautions he has no randomized identification to establish causation.
“we find more licensing, a more licensed workforce, is associated with more joblessness... it's basically significant in all of the past 4 decades.”