Silicon Valley’s Dilemma Over Credentials

Silicon Valley has been heavily derided by the media these past few months, on everything from Google executives forcing senior citizens out of their homes and founders accused of sexually harassing women to startups stealing reservations and parking spots. Despite the shrill news around “JerkTech,” I believe that Silicon Valley on the whole is fundamentally a decent place, one that tries to create opportunities for the world rather than diminish them.

To see this, take a look at our industry’s efforts to make the acquisition of credentials more open and democratic, particularly around those in education. Today, the most important credentials for new graduates continue to be college name, GPA, major, and internship employer names. Yet, many of these credentials are extremely difficult to afford. Average university tuition has skyrocketed over the past few decades, and internships in most industries are rarely paid (except in our industry, where even high school students can earn thousands per month).

Some notable people are trying to change this situation. Peter Thiel, who has warned of a “college bubble” due to the cost of higher education, has sponsored the eponymous Thiel Fellowship, which provides a set of twenty students under the age of 20 annually with $100,000 to drop out of college and work on their ideas.

Thiel is hardly the only one hoping to transform our society’s obsession with college degrees. The entire premise of massively-open online courses (MOOCs) like Coursera and Udacity is to uncouple the knowledge gained from courses from the slips of paper offered at graduation. Both now offer their own credentialing systems, which over time could offer an alternative credentialing model that would be more affordable for workers.

Unfortunately, Silicon Valley’s efforts here have mostly centered around changing the credentials themselves rather than reducing demand for them in the first place. And that is the serious conundrum facing entrepreneurs hoping to build a more meritocratic and open labor market.

Productivity growth has made many people irrelevant in our economy, and will continue to do so well into the future. That means fewer good jobs, even though the number of college graduates is increasing.

Competition is rising for the best jobs in the economy as more people apply for a limited set of opportunities. As our technologies have improved the nation’s productivity, there has been a decrease in good jobs available to workers (this data is discussed more below). Yet, the human population continues to grow in much of the world, with the United States alone projected to reach 400 million people by 2051, up from around 300 million today according to the Census Bureau.

With such a labor environment, credentials are becoming even more crucial in differentiating talent, the exact opposite goal we are striving towards today. While it is hardly desirable to reverse our efforts around productivity, if we hope to change the way credentials are used in our economy, Silicon Valley should emphasize more of its resources around rebalancing the labor market rather than merely changing the pieces of paper used.

The Productivity Crisis

In a short book from 2011, Gallup chairman Jim Clifton wrote that “the coming world war is an all-out global war for good jobs.” He heavily underlined the importance of “good jobs,” which can be defined as jobs that provide a stable, regular income in exchange for work. The challenge for policymakers is that good jobs are declining as a percentage of the overall workforce, an effect largely driven by growing productivity from technology and economies of scale.

While Washington often fixates on the unemployment rate as the barometer for the country’s economic health, that statistic is heavily misleading, since it assumes that any work is better than no work. As can be seen in data from the Bureau of Labor Statistics, the jobs being created by our economy are increasingly “bad” jobs – those jobs that are highly unstable, provide few benefits, and have little hope for advancement.

This is a productivity crisis, brought on by the incredible advancement of technology by Silicon Valley. In almost any context, greater productivity is something to rejoice. With less labor required to build outputs, we can enjoy more of the fruits of our economy without increasing our workloads.

The challenge has been that those fruits have not been shared well, but rather have been increasingly concentrated in a narrow elite of the population. Middle-class jobs are disappearing and are being replaced almost entirely with lower-class jobs. Even with taxes and income-transfer programs like Social Security and TANF, we don’t have robust mechanisms to guarantee that workers left out of the elite workforce still have the means to have a decent and stable lifestyle.

Silicon Valley seems to be more reflective about this topic these days. Just this week, Google’s chief executive Larry Page talked with venture capitalist Vinod Khosla about the future of the 40-hour workweek. Page, quoted in Mashable, said “… the idea that everyone needs to work frantically to meet people’s needs is just not true.” He noted that while work was necessary for humans to “feel needed,” the solution to our current employment crisis was to reduce the workweek rather than just eliminate jobs.

There are certainly ways that technology can be used to make discovering work easier, particularly for those who want to stitch together several part-time jobs into a full work schedule. As I speculated earlier this year about what such a labor protocol might look like, it seems that we have an opportunity in Silicon Valley to radically redefine how work gets done.

But those hopes and opportunities don’t change the reality facing families across America. Productivity growth has made many people irrelevant in our economy, and will continue to do so well into the future. That means fewer good jobs, even though the number of college graduates is increasing.

The Everlasting And Narrow Power Of Credentials

Silicon Valley has a complicated relationship with credentialing, often arguing for the democratic elitism of “meritocracy,” emphasizing proven results and one’s actual ability to build products. In reality, it still clings to the same credentialing system used by the rest of the economy, where brand names, particularly those in education, have disproportionate power in the labor market.

That has always been part of the irony with Peter Thiel, who holds two degrees from Stanford yet encourages others to drop out through the Thiel Fellowship. It’s also doubly ironic since many of these dropouts end up returning to school, as we noted earlier this week.

The problem for Silicon Valley is that credentials are not just used as devices for recruiting, but instead as a means of social proof.

Let’s talk a bit about credentials themselves, a concept I have thrown around a bit without adequately defining. A credential in the labor market is some sort of heuristic about the future performance of a worker. Since an employer has not properly vetted a potential employee through actual job performance (interns are an obvious exception), it must rely on past data to predict success in a role.

Ideally, an employer would spend significant time poring over a candidate’s records, reading their previous output of college essays and GitHub contributions. In reality, such search costs are hard to bear, particularly since keen competition for jobs has increased the size of applicant pools.

The best credentials obviously provide high signal. These credentials provide a way to segment a group of job applicants into two groups, one with high potential for success, and the other with low or lower potential. The smaller and more accurate the identification of the first group is, the better. If a certain credential could screen all but one candidate out of the applicant pool who had a guaranteed chance of success at a new role, it would be perfect.

But signal is just one factor that makes a great credential. Recognition matters a lot too. Presenting a research paper at a specific computer science conference might be incredibly prestigious, but if no recruiter understands the signal inherent in such a talk, then it is mostly useless from a labor market perspective. There is a balance then, between having too few people with a credential and making it irrelevant, and having too many people with a certain credential making it low signal.

To get a sense of what these numbers look like, take a look at the enrollment size of top universities like Harvard and Stanford. Nearly all top schools ranked by US News hover between 1,000 and 1,700 students per class, with some notable exceptions like Caltech having fewer students. In fact, it is only at #16 with Cornell and #20 with the University of California – Berkeley that we see entering class sizes significantly deviating from this rule.

In addition to signal and recognition, a credential must also be simple to understand and widely comparable. Having an engineering team read over the source code of a candidate is probably one of the best ways to see a potential engineer’s performance. But with limited time, that mass of data is often projected down to a single number like GPA or number of stars on GitHub. This also allows candidates to be compared in an ostensibly more “objective” manner even if they have widely different backgrounds.

If these three factors were all that mattered, then startups would have a great shot at transforming credentials. Much like startups are disrupting FICO Scores and the measurement of credit worthiness with better data, it would seem obvious that better data could change the way recruiting works. Imagine taking advantage of all the data available from a candidate’s portfolio and using machine learning to judge a candidate on a small scale of numbers. There are startups like Gild which are doing exactly this.

The value of credentials will automatically decrease if our labor markets are more balanced between workers and corporations.

The problem for Silicon Valley is that credentials are not just used as devices for recruiting, but instead as a means of social proof. Companies in industries as diverse as Google, McKinsey, and Goldman Sachs want to select graduates with certain credentials to maintain a certain “culture” or “fit.” In fact, some industries like management consulting are entirely predicated on hiring for these sorts of credentials. Clients often care about the credentials of the professionals they employ, and they are hardly going to look into constructed metrics by startups to judge their lawyers, bankers, or investors.

Given all of these factors, companies are going to continue to demand candidates with certain credentials, because it makes their job easier and it ensures a certain culture that may be more difficult to develop through other means.

Solving Credentials On The Demand Side

So, we have fewer good jobs being created due to our productivity growth, and we know that companies desire candidates with better credentials than those with worse credentials, because it is simpler and safer. Thus, in a tough labor market, credentials are crucial for workers hoping to secure a good job in our economy. As competition increases for a limited number of slots, the standard required to secure a particular position also rises, which means that certain credentials become even more important just to get employment.

This basic analysis is known intrinsically by most middle-class families in the United States, and explains why so many parents send their kids to college, even though the kids may not be ready or interested in such a pathway. Nearly any professional job requires a college degree today, a simple function of the depth of the labor talent and the shallowness of the labor demand.

Given this analysis, I would argue that startups are working on the wrong set of challenges. They should be much more focused on creating good jobs, rather than trying to minimize the use of credentials in employment or transforming those credentials into something more open. The value of credentials will automatically decrease if our labor markets are more balanced between workers and corporations.

We don’t have to cognitively travel far to see this – just take a look at our own industry. As demand for engineers has increased because of Silicon Valley’s booming startup ecosystem, there is now much more interest in looking beyond the typical college degrees and academy companies. The demand for labor is the only value for a startup like Gild, which attempts to discover engineers who might not otherwise show up on a company’s recruiting system, because maybe they lack the typical college credentials or their work on open-source projects is not fully appreciated.

To see the opposite trend, take a look at a professional labor market like legal services. There, the top jobs go almost exclusively to students graduating from the “T–14” set of top law schools. The difference in work outcomes is staggering. Take a look at this graph from the National Association of Law Placement. Those entering the top firms will make an income around $160,000. Otherwise, the typical job pays around $50,000, or maybe even hourly these days. Due to the glut of new lawyers, legal credentials have unbelievable power on the job outcome here, unlike in Silicon Valley for engineers.

Credentials are an ingrained part of how we judge candidates, and startups trying to change this culture are facing a Herculean task trying to persuade human resources professionals to adopt a new paradigm. But we do have the ability to make credentials less useful if we can provide better income opportunities and increase the competition for talent.

Photo by Flazingo used under a Creative Commons 2.0 license.