On Skills and Schooling
The notion that America’s education system is failing to equip the workers of tomorrow is at least as old as the seminal 1983 government report A Nation at Risk, which memorably quipped, “If an unfriendly foreign power had attempted to impose on America the mediocre educational performance that exists today, we might well have viewed it as an act of war.” Today’s CEOs regularly blame schools and colleges for their difficulties in finding adequately prepared employees. The complaint shows up in survey after survey, as Cappelli shows in his book, and it is substantially more common among American employers than their peers in most other developed and developing economies.
But do these surveys “show that the United States is among the world leaders in skills gaps,” Cappelli asks, “or simply in employer whining and easy media acceptance of employer complaints?”
He thinks a body of lesser-reported studies contains the answer. “If you look at the studies of hiring managers and what they want, they’re not complaining about academic skills,” Cappelli says. “You hear the business spokespeople saying this, but the actual hiring managers are not saying this now. And in fact they’ve never, in modern times, said that.”
In other words, there’s a disconnect between the people who actually do the hiring and the C-suite executives who get asked about it.
“The spokespeople at the very top don’t know what’s going on,” Cappelli argues. “But there’s a popular story that explains what’s going on, and that story is: schools are failing.”
Cappelli reckons there are two main reasons this story has staying power. “One reason it’s popular is that it allows you to advocate for school reform, which sounds like a socially desirable sort of thing to do. And so all these different groups like this story [too], because it works for the people who like to think that they’re interested in reforming society and improving education. What’s not to like about that?
“It also works for employers who are saying: This is not our problem. It’s a problem the public sector ought to solve for us. They ought to get better at providing the things that we need, and then we don’t have to do anything—we don’t have to train, we don’t have to adjust our expectations.”
The problem with the story, Cappelli says, “is there’s no evidence for any part of it.” And what’s more, it’s hurting the very companies that espouse it.
More on that second point later. Before we get there, what are hiring managers complaining about? For the most part, Cappelli says, work attitudes and lack of on-the-job experience—neither of which can be fairly laid on schools. When you dig deeper into the surveys of talent shortages, a similar theme emerges. Cappelli notes that the top 10 hardest-to-fill jobs globally in 2011, according to a well-regarded survey from the Manpower Group, included positions such as laborer (“about as unskilled a job category as we can get”), production operator (“a factory job, typically semiskilled”), technician and skilled trade (“those skills are learned largely on the job”), and sales rep.
Of the top 10, he reckons that engineers and accounting/finance workers learn their skills largely in college classrooms. The same is probably true now of IT staff—but wasn’t always, which brings us to another of Cappelli’s insights.
The Home Depot Syndrome
In a 2011 op-ed article for The Wall Street Journal, Cappelli remarked on a telling statistic from the Silicon Valley tech boom of the 1990s: only 10 percent of the people in IT jobs had IT-related degrees. But a lot of the same people would probably have a hard time landing similar jobs today, because employers have increasingly adopted what Cappelli calls “a Home Depot view of the hiring process, in which filling a job vacancy is seen as akin to replacing a part in a washing machine.
“We go down to the store to get that part,” he explains, “and once we find it, we put it in place and get the machine going again. Like a replacement part, job requirements have very precise specifications. Job candidates must fit them perfectly or the job won’t be filled and business can’t operate.”
The problem with this approach, he says, is that “unlike a machine part, no perfect fit exists between applicants and job requirements.” In his book, Cappelli cites a series of studies by the National Institute of Economic and Social Research in London that investigated how companies “making almost identical products but operating in different countries got their work done.” US firms used more engineers and unskilled workers, for instance, than German firms, which relied more heavily on skilled craftsmen. In other words, there are multiple ways—and multiple kinds of employees a company might rely on—to accomplish a given task. Narrowly drawn job criteria may be a sign that a company is ignoring possibilities for alternative, and perhaps even more effective, operational strategies. To the extent that they slow down the hiring process, a company might stand to benefit from adopting a more flexible approach to hiring.
There’s a simple lesson there. “When employers have a vacancy to fill,” Cappelli writes, “they have many options for filling it.” That sounds like common sense, but it runs us smack into a new reality of the contemporary hiring process: common sense is exactly what software-based hiring systems lack.
The Software Ate My Job Application
Remember the story about the company that couldn’t find a garden-variety engineer out of a deluge of 25,000 applicants? Is such a thing really possible?
More possible than you might think, Cappelli says.
If you’ve applied for a job through a web-based interface that requests everything from your most recent job title to your Excel proficiency to your grades in high school, you may have discovered this the hard way. In many common applicant-screening programs, each question functions like a gate. If your answer to any of them doesn’t fit the precise criteria set by the company, that gate slams shut and the software deems you unqualified. Alternatively, some systems rely on automated keyword searches of resumes, rejecting those that don’t produce exact matches with job descriptions that are by turns overly detailed or hopelessly vague.
In this manner, a company can hamstring its own applicant search pretty quickly—even if its job posting draws 25,000 responses. “All one needs for that to happen are 14 requirements in the model just discussed, many fewer if some hurdles are highly specific,” Cappelli notes. And as crazy as it sounds, mundane choices like how a job-seeker formats his resume can also hurt his chances.
(For Cappelli’s advice on “beating” the software, see sidebar.)
Not everyone shares Cappelli’s view of software-based hiring. Gary Truhlar WG’74 is the executive director of human resources at Penn, which is the largest private employer in Philadelphia.
“When I first came to Penn, we would post job opportunities on seven bulletin boards,” Truhlar recalls. “They were xeroxed, behind date, and physically posted on bulletin boards—so if you weren’t on campus, you didn’t know about those jobs. We would run a small subset in the Inquirer, but that’s very expensive. So the good side of the technology, posting on the web, is that as soon as HR knows about the job, the world knows about the job.”
That translates into a broader and more diverse applicant pool. Job postings at Penn frequently attract 300 or more candidates. The dramatic expansion of applicant pools has presumably led to a more talented workforce.
“The downside,” as Truhlar points out, “is that anybody can apply for anything. So, Peter’s work notwithstanding, you’ll have a secretary applying to be a vice president.”
Penn uses a web-based applicant management system to handle the volume—but, Truhlar stresses, not one that relies on keyword searches or binary gates that stack the deck toward rejection. “We are concerned that applicants are not eliminated because of the system limitations created by certain application providers,” he notes. So the system Penn uses—called PeopleAdmin—is designed to assign scores to applicants, not simply render them qualified or unqualified. So even if no one scores 100, a hiring officer can use the scores as a starting point for making more nuanced judgments.
The notion that a firm hiring engineers would let 25,000 people apply without finding one “stretches my imagination,” Truhlar says.
“We just filled a data-security and privacy position,” he adds by way of example. “And you can kind of do the same thing. You can say, Do you have five years of experience in the data security business? Do you have Microsoft certification? And that’s terrific—that’s probably absolutely what you need. Well, then you can make the mistake of saying, I need Microsoft x.3.7, and I need Symantec version 8.2.1—and that’s how you get into a situation where 25,000 people didn’t qualify. And I would say those HR people are useless, if that’s all they’re doing is a keyword search and they can’t find anybody.”
Cappelli would agree, but perhaps go one step further by noting what he sees as a major reason HR departments don’t succeed: the departments themselves have been driven to the brink of extinction.
Human Resources in Bad Decline
“One of the things which is a great puzzle,” Cappelli says, “is that 50 years ago in the US, most big employers were much more sophisticated about hiring than they are now. So they’ve actually gotten worse at this.”
The body of knowledge about what predicts job performance goes back to World War I, Cappelli says, “and it’s just being systematically ignored.”
Take the case of high-school GPA. “We know something about this!” he says. It “predicts nothing about your job performance—especially 30 years later. Why are they bothering to ask that? We know it doesn’t work.”
He has a ready answer to that question: “the gutting of the human-resource function—cutting people out, cutting staff out. The people who were trained in this stuff, and used to know it, are all gone now.”
This is arguably the product of systemic changes to the US labor market that gained pace in the 1980s. As high unemployment rates from the late 1970s persisted into the following decade, companies increasingly had at their disposal a labor pool that was more experienced than ever before.
“A generation ago you had to hire out of college and train people,” Cappelli explains, singling out firms like IBM and General Electric, which aimed to retain employees for life. “Nobody hopped jobs. But as companies started to lay people off, and there were lots of skilled people around and lots of white-collar workers, lots of managers, you could sort of hire whoever you wanted. It wasn’t very hard. So you maybe didn’t have to be that sophisticated at recruiting and selection” anymore.
Some of the pressures that touched off the cycle of layoffs are well known, like increased global competition, and the perceived need for companies to respond more nimbly to changing consumer preferences (which could make replacing employees seem more efficient than taking the time to train them). But there are other salient factors that are underappreciated. Cappelli outlined one in a 1997 volume titled Changes at Work, in a chapter co-written with Wharton’s Michael Useem, now the William and Jacalyn Egan Professor of Management: the growing concentration of stock ownership among large institutional holders, like pension funds.
Traditionally, such investors expressed dissatisfaction with a company’s performance by simply selling their stock. But with the rise of indexed holdings—in which funds are invested according to preset formulas (for instance, mirroring the composition of the S&P 500)—and the sheer size of share blocks under ownership, it became harder for big institutional investors to shed unwanted stock positions. Consequently, they instead began pressuring companies to change, with a narrow focus on driving up share performance over shorter time periods.
“Institutional holders were now quicker to insist that the company find new strategies and structures to produce,” Cappelli and Useem noted. “And the formulas they found concentrated on restructuring the companies and slashing jobs. One study of share-price reactions to company layoff announcements from 1979 to 1987 illustrates the thrust of the investors’ message for would-be restructurers. In the days immediately following layoff announcements announced as part of general restructurings, stock prices rose an average 4 percent.”
Concurrently, the ranks of experienced job-seekers swelled, making decisions to fire still easier to justify, since the labor market was flush with talent. This dynamic persists today—with a perverse twist.