The worst thing about tech bubbles isn't what you may think

You may recall how the last tech bubble 15 years ago resulted in staggering market losses, numerous failed start-ups and increasing IT unemployment. Less noticed was the bubble's eerie correlation to undergraduate enrollments in computer science.

As the NASDAQ increased, so did computer science and computer engineering enrollments. They were almost in heartrate synch, like E.T. and Elliott (in the blockbuster movie). When the NASDAQ lost half its value in 2000, within a few years computer science enrollments also fell about 50%. Today, enrollments are again spiking right along with the stock market.

The massive valuations of the latest Internet sweethearts has some seeing a bubble in the making. These high valuations include Uber at $50 billion, similar to FedEx, Salesforce, Caterpillar and Target; Airbnb at $25 billion, similar to utility Pacific Gas & Electric and Netflix; and Snapchat's $15 billion, which is roughly equal to Whirlpool and Tyson Foods.

Investors aren't the only ones who get hurt when a bubble pops.

IT employment, excluding tech manufacturing, at 3.527 million in 2001, declined about 5% or by nearly 200,000 jobs in the following year, according to data by TechServe Alliance, an industry group. IT employment didn't regain its 2001 level until 2005.

Meantime, the bubble also crushed computer science enrollments. In March 2000, the NASDAQ composite index reached a historic high of 5,048, at just about the same time undergrad computer science enrollments hit a peak of nearly 24,000 students at Ph.D.-granting institutions in the U.S. and Canada, according to data collected by the Computing Research Association in its most recent annual Taulbee Survey. (There are many more computing science students enrolled at non-Ph.D. schools.)

By 2005, computer science enrollments had halved, declining to just over 12,000. Remarkably, it has taken nearly 15 years to reach the earlier enrollment peak.

On July 17, the NASDAQ hit its highest point since 2000, reaching a composite index of 5,210. In 2014, computer science undergrad enrollments reached nearly, 24,000, almost equal to the 2000 high. They may well be past it now.

"It's often difficult to recognize a bubble while you're in it, as unreasonable optimism and speculative greed lead to the belief that a 'new paradigm' will validate wildly aggressive projections," said Bruce Bachenheimer, clinical professor of management at Pace University and executive director of its Entrepreneurship Lab. "It certainly appears that certain sectors of the market are due for a major correction," he said.

The experience of 2000 dot-com bubble is still fresh for many, and while they see risks and reasons for caution, there are also clear differences between then and now.

Tim Herbert, the senior vice president for research at CompTIA, says the investment landscape today "is still well short of the dot-com bubble days of 2000." The number of IPOs is nearly 70% lower than they were 15 years ago, and among those companies going public, "the trend appears to be towards firms that have a longer track record, solid business fundamentals and earnings." There are also far fewer "mom and pop" investors pouring money into individual stocks, he said.

Herbert still sees some reason for caution. The uptick in VC funding over the last eight quarters "may suggest there are more dollars chasing fewer quality deals." Funding for software firms may soon match the 2000 peak, and those investments also correspond to the rise in cloud computing, he notes. But some valuations "have been pushed to levels hard to justify."

One area of broad agreement is the reduced cost of a start-up. The risks may not be as great as they once were.

New technologies including open source software and cloud computing "have dramatically" reduced the costs of starting a business, said Patrick Gallagher, general partner of VC firm CrunchFund. A start-up that might have needed $5 million to $10 million in funding in 2000 can be established with $1 million or $2 million today, he said.

In the dot-com era, companies could launch and go public "all within 12 months," said Gallagher, with employees of these firms selling their shares. "It just kind of created this unsustainable market because these companies themselves weren't real companies," he said.

Another difference: In the dot-com bubble, firms without revenue were being valued on metrics that had no substance, such as eyeballs. Today, the firms that are getting high valuations are earning revenue and aren't going away, said Glen Rieger, general partner at NewSpring Capital.

The 2000 bubble wasn't a complete disaster. Many of today's firms -- such Amazon, Google and Salesforce -- emerged and have thrived. But there is no escaping the bubble's long-tail impact on undergrad student enrollment, which followed the market down.

Enrollments in computer science at the undergraduate level have in the past "been pretty correlated with the state of the tech industry," said Peter Harsha, the director of government affairs at the Computing Research Association.

The enrollment survey only tracks numbers, but there is a belief that the present surge in computer science enrollments "feels different than the previous peaks," said Harsha. For one thing, there's a sense that "students are engaging in computing because they recognize that it's becoming increasingly required for success in other disciplines as well."

Computing science enrollments will still be cyclical, but this idea -- that computing science is needed today in almost every endeavor ---will mean that "those cycles won't be tied so tightly to the health of the tech industry but rather the health of the economy as a whole," said Harsha.

If we are in a bubble, the impact on computer science enrollments -- similar to what happened in 2000 -- won't be immediate. "The decline in enrollments typically lags about two years the decline in the industry and employment," said Hal Salzman, a professor of planning and public policy at Rutgers University, who studies the science and engineering workforce. "So if this is a bubble, students would be smart to anticipate it in their choice of major."


Patrick Thibodeau

Zur Startseite