Tuesday, June 16, 2015
Tom Friedman's latest New York Times column uses the labor market for executive assistants and executive secretaries to illustrate dubious claims about credentialing and over-education. Friedman argues that since most current executive assistants and executive secretaries don't have bachelor's degrees, employers should not try to upgrade the workforce by hiring new executive assistants and secretaries with bachelor's degrees. After all, executive assistants without bachelor's degrees can do the job, so who needs a bachelor's degree?
The problems with this reasoning should be obvious.
First, education is only one of many factors that are valued in the labor market. Some individuals who are smart, hardworking, personable, physically attractive, or fortunate, but have limited education, will inevitably be as successful or more successful than other individuals who are highly educated but less gifted in other respects. This does not in any way challenge the extremely strong evidence that a bachelor's degree can improve labor market outcomes. It simply means that we are dealing with a heterogeneous population.
If two homogenous groups who were initially equally strong on non-education factors were given different amounts of education, the more educated group would typically be more successful in the labor market. Labor economists who have studied identical twins routinely find that twins with more education are more successful than their less educated counterparts. When labor economists control for unobserved heterogeneity within education levels using fixed effects models rather than OLS regression, "over-education" effects on earnings diminish or disappear. In other words, highly educated folks who are about as successful as those with less education--and end up in the same occupations as the less educated--tend to be weak on factors other than level of education. But even within occupations that combine the worst of the more-educated with the best of the less-educated, those who are more educated still tend to earn more. Since profit-maximizing employers are not in the habit of handing out money for nothing, this suggests that the more educated are better at their jobs.
In sum, education many not always be enough to make you more successful than your neighbor or coworker, but it can make you more successful than a less educated version of yourself.
Second, the fact that something was "good enough" at some point in the past does not mean it is good enough today. Rising standards typically involve both increases in quality and commensurate increases in cost. In inflation adjusted terms, the average new car today costs about 10 times as much as a Ford Model-T in the late 1920s. But the average new car is faster, safer, more reliable, and easier to operate. Similarly, as education increases, so does the productivity of labor and the cost of labor--wages or earnings. Highly educated workers today are far more productive than their counterparts decades ago, and as a result, they earn more.
It is interesting that Friedman chose executive assistants and executive secretaries--a field where most workers have less than a bachelor's degree--as an example of supposed "over-education." According to the Bureau of Labor Statistics Occupational Employment Statistics, employment of executive assistants and executive secretaries is collapsing. Employment fell by more than half between 2007 and 2014, from over 1,500,000 workers to barely more than 700,000. In other words, the level of education that most executive assistants and secretaries had in 2007 was not enough to make it in the labor market of 2014.
Among secretaries, those with higher levels of education still earn more than their less educated counterparts after controlling for race. Employer hiring priorities cited by Friedman suggest that those who are more educated are more likely to keep their jobs or find new ones.
This is consistent with general trends in the labor market. Low and middle skill workers with limited educations are the hardest hit by automation, outsourcing and layoffs, while their more educated counterparts are navigating the recession and changes in the labor market more successfully. (During the 2007-2014 period, employment of a group of highly educated workers, lawyers--supposedly the victims of job-destroying structural change--continued to grow faster than overall employment).
For another angle on Friedman's column, readers may be interested in Frank Pasquale's critique. Pasquale discusses apparent bias in the New York Times' Higher Education coverage and argues that as newspapers struggle to adapt to a world replete with free online content and greater competition for advertising dollars, business priorities may be overriding traditional news values. Given the nearly 20 percent decline in employment for reporters and correspondents between 2007 and 2014, journalism does appear to be under serious financial pressure.
Monday, June 15, 2015
A shorthand approach sometimes used to compare the cost and benefits of higher education—comparing student loan balances at graduation to first year earnings—can be seriously misleading. The implication of this approach is that student loans have to be repaid in full shortly after graduation, and that graduates’ low initial earnings will persist for the rest of their lives.
This is an apples to oranges comparison. An investment in education pays dividends throughout one’s life. First-year earnings are one small, unrepresentative, slice of lifetime earnings. Comparing a lifetime investment to one year of expected returns on it feeds ignorance about how student loans and lifetime earnings actually work. It thus risks misleading prospective students into making financially disastrous decisions to underinvest in education.
Student loans are meant to solve a specific problem—the costs of education come as a series of large upfront payments for tuition and living expenses, while the benefits accrue later in life in the form of higher earnings. Except for the minority of students who are fortunate enough to have rich and generous parents who cover their tuition, students generally have two options—save or borrow.
Saving is inefficient because it requires students to work for many years with a lower level of education and for much lower wages, and to complete their degrees much later in life. Completing a given level of education earlier helps maximize the number of years of expected higher earnings with a higher level of education. Borrowing to invest in education is therefore more efficient than saving to invest in education. Some of the benefits of financing accrue to the student borrower in the form of higher lifetime earnings compared to saving, and some of the benefits accrue to the lender in the form of interest and fees. Another approach to financing higher education—more popular in Europe, Australia, and Canada than the United States—features higher public spending and higher tax burdens, sometimes with a tax-like percent-of-earnings fee explicitly tied to university education. The social democratic approach, like the U.S. approach, involves providing something of value up front in return for a fraction of graduates’ incomes later.
Student loans enable students to pay for their own education by converting the cash flows associated with investment in education from large upfront payments into a series of much smaller payments spread out over time. Ideally, these payments should closely match the timing of the benefits of education—that is, the timing of the boost to earnings from education.
Because the benefits of education accrue over the course of a career—perhaps 40 years or more—and earnings typically do not peak until middle age, the costs of education should ideally also be spread over a similar time frame.
The prospect of high payments needed to pay back loans very quickly ex-ante could cause prospective students to underinvest in education. As life expectancy and career length increase, so should initial investment in education.
If this goal of matching the timing of cash flows is accomplished, then at every point in time, with more education, students will have more cash at their disposal. The boost to earnings from education will more than cover student loan debt service payments, and the initial borrowing will enable students to maintain a decent lifestyle while pursuing studies instead of working full time. (For a discussion of the advantages of leveraged investments early in life, see Ayers & Naelbuff).
That is one important reason why federal student loans can be repaid over 25 to 30* years (so-called “extended” repayment). Plans are available under which monthly payments start low and increase over time to match the typical trajectory of lifetime earnings (“graduated” or “graduated extended” repayment), or in which payments dynamically adjust up and down with actual borrower earnings (if earnings fall below a certain level) to better match cash flows (“income-contingent” or “income-based” repayment).
Because these extended and income adjusted plans are better tailored to the purpose of student loans—matching positive and negative cash flows—one of these plans should be the default option for student borrowers instead of the now “standard” 10-year repayment period. 10-years to pay for an education that provides benefits over 40 years makes little sense. For law graduates, real earnings typically continue to grow for 30 years after graduation.
* Consolidated loans can be repaid over 30 years, but some consolidated loans may not be eligible for income based repayment plans.
Paying loans back slower typically will not affect the economic value of education, notwithstanding the fact that nominal interest payments will increase. Paying loans back faster or slower typically will not affect the economic value of education as long as two conditions are met:
- Interest rates remain unchanged regardless of whether a loan is repaid over 10 or 30 years (this is the case for federal student loans, but not for mortgages or most other debt instruments most of the time)
- The interest rate on student loans is appropriate, in that it matches up with default and loss risk levels for lenders, the opportunity cost of capital, and time preferences.
If condition 2 holds, then the interest rate will equal the discount rate which is used to convert cash flows occurring at different points in time into the same currency so that they can be compared. If the discount rate is 6 percent, then there is no valuation difference between paying $1,000 today or paying $1,060 one year from now, just as there is no difference between paying one U.S. Dollar or the equivalent in Euro cents. If students choose to refinance or pay their loans back faster than they are legally required to repay them, this suggests that the interest rate on student loans is too high.
Saturday, June 13, 2015
It's been a pleasure and a privilege to teach such talented young men and women, and I am sure I speak for all of my colleagues in wishing you much professional success and personal happiness in the years ahead.
June 13, 2015 | Permalink
Friday, June 12, 2015
The Department of Education has been overcharging low-risk professional school students for federal student loans (relative to the market rate) while keeping rates low for undergraduates who are far more likely to default. (For previous coverage, see here, here and here).
Bloomberg BNA's Bankruptcy Reporter describes the predictable consequences of this politically driven mispricing: Professional graduates are refinancing into less expensive private loans and removing themselves from the government's risk pool.
There is a simple solution that will shut down what Bloomberg describes as an "exodus of top borrowers" while preserving student lending profits for the benefit of taxpayers. The government should charge low risk graduate students less.
Update, June 13, 2015: Jordan Weissmann at Slate covers the story.
Monday, June 8, 2015
1. Lucian Bebchuk (Harvard) (230,377 downloads, 172 papers)
2. Daniel Solove (George Washington) (229,918 downloads, 41 papers)*
3. Cass Sunstein (Harvard) (205,141 downloads, 189 papers)
4. Mark Lemley (Stanford) (161,607 downloads, 141 papers)
5. Bernard Black (Northwestern) (161,459 downloads, 138 papers)
6. Stephen Bainbridge (UCLA) (111,432 downloads, 95 papers)
7. Brian Leiter (Chicago) (103,669 downloads, 59 papers)
8. Dan Kahan (Yale) (95,120 downloads, 54 papers)
9. Eric Posner (Chicago) (92,878 downloads, 122 papers)
10. Orin Kerr (George Washington) (89,492 downloads, 49 papers)
*A single paper accounts for nearly two-thirds of Prof. Solove's downloads!
Thursday, June 4, 2015
Monday, June 1, 2015
Thursday, May 28, 2015
1. Cass Sunstein (Harvard) (28,599 downloads, 24 new papers)
2. Dan Kahan (Yale) (18,796 downloads, 5 new papers)
3. Daniel Solove (George Washington) (18,503 downloads, 2 new papers)
4. Mark Lemley (Stanford) (14,973 downloads, 8 new papers)
5. Lucian Bebchuk (Harvard) (13,940 downloads, 0 new papers)
6. Orin Kerr (George Washington) (12,254 downloads, 4 new papers)
7. Brian Leiter (Chicago) (12,097 downloads, 9 new papers)
8. Bernard Black (Northwestern) (10,561 downloads, 5 new papers)
9. Jeremy Waldron (NYU) (8,214 downloads, 6 new papers)
10. Tim Wu (Columbia) (8,158 downloads, 2 new papers)
And given how close to the top ten, I should note that my colleague Eric Posner had 8,065 downloads and six new papers in the last 12 months.
As the cases of Solove and Bebchuk show, "oldies but goodies" can keep the downloads pouring in!
Tuesday, May 26, 2015
Monday, May 25, 2015
...while undertaking additional cost-cutting measures. It appears the School enjoys some strong support in the local Charleston community.