About 10 percent of Americans today don't have access to high-speed Internet service. The rest of us are pretty much stuck in the granny lane—on average, we get broadband speeds of less than 5 megabits per second, 10 to 20 times slower than what people in many other countries enjoy. Derek Turner, research director for the public policy group Free Press, dreams about America becoming a "broadband utopia." In Turner's paradise, you'd be able to order up a fast connection no matter where you lived. And not just that—several broadband providers would compete with one another to bring super-fast service to your door, a dynamic that would keep prices low and speeds very high (100 MBps downloads and uploads, a file-trading gamer's promised land).
In a policy paper that Free Press put out in December, Turner and his colleagues called on the Obama administration to spend $44 billion to realize this dream. The broadband advocates argue that the money would boost short-term economic activity—we'd need tens of thousands of people to produce and maintain fiber-optic cables, routers, and other equipment; to dig trenches and climb poles to install the new broadband lines; to staff customer service and billing centers; and to train everyone to use the new stuff. The long-run effects of a national broadband plan are even rosier. More than any other investment, Free Press argues, Internet lines would stimulate activity broadly across the American economy, fostering innovation and new jobs in education, health care, retail, and high-tech businesses.
That sounds great—sign me up! But wait a second … here's Don Detmer, president of the American Medical Informatics Association. In an open letter to Obama shortly before the inauguration, Detmer called on the new administration to spend $10 billion a year for five years to create electronic medical records, a huge project that would require the training and hiring of tens of thousands of new health care and technology workers. Not only would the plan create new jobs, Detmer says, it would also reduce costs by making medicine more efficient. Plus, the investment would improve our health—electronic records would allow for more advanced medical research and significantly reduce errors. (Doctors' sloppy handwriting supposedly kills more than 7,000 patients a year.)
OK, now I'm confused. Should we spend stimulus money on building a broadband utopia or on transforming health care? Or both? Or maybe, as Thomas Friedman has argued, we ought to build advanced batteries, hybrid drivetrains, and other environmentally friendly technologies. That sounds great, too, right? But what about a completely redesigned "smart" electricity grid that would be able to handle a new generation of plug-in cars, fuel cells, and other as-yet-unimagined power-generation technologies—what some people have called the next Internet. Let's hear it, American taxpayers: Which of these things, if any, do we want to fund?
The two stimulus plans working through Congress outline far fewer funds for high-tech projects than their advocates would like. The House and the Senate are calling for around $6 billion or $7 billion to fund broadband infrastructure, with much of that money going to rural areas—a far cry from the $44 billion Free Press asked for. Congress seems to be settling on about $20 billion to $25 billion to improve health care technology—about half or a third of what estimates say a transition to e-health records would cost.
The problem with calling for any government funding of technology is that the future always sounds terrific. Who doesn't want cheap Internet everywhere, an end to medical errors, and an electric system that could change the way we drive? Sketched out like this—a series of plans that promises radical advancements after a relatively small investment of resources—it seems crazy not to sign up for every one of these ideas. After all, the U.S. government has played a huge role in the inception of nearly every modern innovation we enjoy today. Government research grants were present at the creation of microprocessors, databases, the graphical user interface, video games, the Internet, and the World Wide Web, among many other great things. (See this research report.)
But spending on tech can be very tricky. Advocates for a high-tech stimulus aren't calling for much research money. Instead they're arguing for spending at a more advanced stage of development—they envision the government sponsoring the creation and deployment of ready-to-use technology. And we're all familiar with spectacular government-funded tech failures at that stage—think missile defense, the terminally broken computer systems built for the IRS and the FBI, and the Census Bureau's stalled effort to automate its data collection. The government is not alone in tech incompetence; high-tech companies themselves are regularly blindsided by the future. Look at Sony: Why did the company that brought us the Walkman fail to anticipate the iPod? Why did the company that brought us the PlayStation and the PS2 fail with the PS3? Apple built the first mainstream operating system with a graphical user interface—and then lost that business to Microsoft, which saw that the big money was in the OS, not computer hardware. These failures illustrate the profound difficulty of constructing any kind of tech stimulus package—the past seems to offer little guidance on what people will want tomorrow. But they also suggest a way for President Obama to avoid such pitfalls: He should use the stimulus money to set up something like a government-sponsored venture capital fund. The administration could give out a little bit of money to give a boost to a lot of great ideas, then continue to fund only those ideas that succeed.
After all, the risks in these plans are clear. Over the last few years, a handful of American cities have built systems to provide low-cost wireless Internet service to their citizens. Most haven't taken off. Would an effort to bring broadband infrastructure to rural areas suffer the same fate—what if we built it and no one signed up? The push for electronic health records could also be a boondoggle: What if patients, wary of the privacy safeguards, balk at having their medical histories computerized? What if hospital workers were slow to learn how to use the new systems? What if the software at different hospitals or different parts of the country were incompatible? (See the computer systems at the various intelligence agencies.)
One way to avoid such problems would be for government to stay out of tech innovation entirely. Last week I had a long talk with Jim Harper, the director of information policy studies at the Cato Institute, the libertarian group that opposes any stimulus plan whatsoever. In general, Harper favors tax breaks and decreased regulation to spur private-sector innovation in broadband, health care, and other parts of the economy. When I asked about electronic health records, Harper pointed to the success of OpenTable.com, the site that's brought Web-based restaurant reservations to many cities across the world. OpenTable gives diners a convenient way to make reservations; restaurants looking to attract these customers pay a monthly fee to participate in the service. Harper argues that if we reduce regulation in health care, we'd see an electronic records solution emerge from the private sector, just like we saw a reservation system emerge in the restaurant industry. A health-record startup could enlist a forward-thinking set of hospitals and insurance companies and then give patients some inducement—discounted insurance or drug prices—to computerize their records. If patients show an interest in that plan, it would spread across the country; if not, it would die.
But Harper's ideas sound just as dreamy as those coming from the other side. Detmer points out that half of all money now spent on health care in America is spent by the government—so unless you expect the government to get out of health care entirely (unlikely, and, to many people, unwise), you can't improve medical technology without having the government take part. A similar problem dogs innovation in broadband: A long history of poor telecom policies has left us with just two entities providing fast service—the big phone companies and the big cable companies. Verizon is the only company in America that has large-scale plans to deploy fiber-optic lines to homes—and only 6 million of them. That's why, as Free Press argues, the government is the only entity that can bring truly high-speed broadband to the masses.
As it's constructed now, the stimulus bill would give the administration carte blanche to choose between technological investments. The Senate version of the stimulus bill, for instance, allows the administration to give grants to any number of projects that promise to bring broadband to underserved areas. In order to keep waste to a minimum, President Obama ought to allocate the funds in a way that has proved successful in Silicon Valley: He should act as an angel investor, giving lots of little grants to sponsor different technological approaches to the same problem. For instance, the administration could invest in a dozen or more health-record systems, then see which ones prove most attractive to patients and doctors. In the case of broadband, the government could parcel out a bunch of $10 million research grants to cities that come up with a viable plan to bring fiber-optic lines to their citizens. The handful of cities with the best plans could then be eligible for $100 million in funds to fully build out their ideas.
This funding scheme probably wouldn't work as quickly to stimulate the economy. We'd probably create a lot more jobs in the short run if the government just handed out $20 billion to a single contractor to build a national health-records system. That's how the Census Bureau went about its plan to arm its data-collectors with PDAs rather than clipboards. But guess what—they're still using clipboards.