Even in coal's heyday, Appalachia was still relatively poor and backward. At the time, policy makers blamed its lack of economic development on mountainous inaccessibility. Their solution: End the region's isolation with massive infrastructure projects, most notably a network of four-lane highways that would connect the region to the rest of the country.

So in 1965, President Lyndon Johnson signed the Appalachian Regional Development Act, creating the Appalachian Regional Commission (ARC). Over the subsequent five decades, ARC has spent $27 billion (in 2015 dollars) to build nearly 3,000 miles of the Appalachian Development Highway System that is threaded throughout the mountains.

The highways, constructed along officially designated "Corridors," are splendidly engineered—and largely empty. They utterly failed to spark an economic renaissance. Despite tens of billions in federal money, the "region's performance relative to the national average is similar to its position in the 1960s," reported economists Carl Kitchens and Taylor Jaworski in a 2016 study published by the National Bureau of Economic Research. They calculate that the gigantic transportation investment boosted incomes in the region by just $586 per capita.

Far from being discouraged by this result, policy makers are at it again. This time, they want to drag Appalachia into the 21st century over newly installed information superhighways, known—God help us—as "eCorridors."

Here's the plan: First lace the mountains with high-speed broadband fiber-optic networks to connect the region to opportunities in the outside world. Then train unemployed miners in the art of computer coding. The first step aims to generate new jobs by luring companies to the area; the second is supposed to let people stay put and work.

I grew up as a hillbilly in central Appalachia, on a dairy farm in Washington County, Virginia. Like many folks, I left to seek an education and better opportunities beyond the confines of the Mountain Empire. I returned for a week in June to cruise the mountainous Corridors and meet with some of the people in Eastern Kentucky and Southwestern Virginia who are trying to jumpstart a hillbilly tech revolution. But instead of a burgeoning tech sector fed by glorious new fiber-optic cables, I found pure deja vu: Underutilized, debt-saddled infrastructure projects and an ever-growing number of Appalachians being expensively trained for jobs that are unlikely to show up.

Hope or Hype?

"Silicon Holler: How workforce retraining is bringing tech jobs to Appalachia," blares the headline in TechRepublic. "Can an Appalachian 'Silicon Holler' rise in coal's shadow?" asks Reuters. The Guardian informs us that the fiber-optic cables being built across Kentucky could transform coal country into "a new place on the map the hopeful call 'Silicon Holler.'"

The hype began as far back as 1999, with a project launched by Bristol Virginia Utilities (BVU), the city agency in charge of providing water, sewer, and electricity services to the 17,000 residents of Bristol, Virginia. That year, the utility proposed and the City Council approved a fiber-optic network to connect its eight electric substations and all city offices, including City Hall, public schools, libraries, and the police and fire departments.

That might have been seen as a logical extension for a utility company. But mission creep was inevitable, and in 2002, BVU began deploying a fiber-to-the-home network for residential customers. At the same time, it started to expand its OptiNet broadband network into Southwestern Virginia using revenue bonds, plus grants from the federal and state governments and tobacco settlement money—for a total of $132 million spent. In the end, OptiNet managed to pick up 13,000 customers and get spun off into an independent authority with its own board of directors.

"Everyone knew that broadband would help the economy in the future, but nobody knew how."

Cash inflows from successive government grants enabled OptiNet to function like a Ponzi scheme, masking the fiscal rot at the heart of the enterprise. Eventually in 2013, an audit found extensive misuse of funds—personal trips, bribes, and kickbacks—by board members, officers, and contractors. In 2016, nine people associated with the BVU Authority, including its CEO, chief financial officer, and board chairman, were sent to prison for conspiracy and fraud. The state government's 2016 final report noted that the OptiNet division was operating at a net loss, that this was expected to continue, and that therefore it was unlikely to generate enough cash to pay both the principal and interest owed on $45.5 million in bonds it issued in 2010.

The audit also found that the BVU Authority used an improper methodology to account for and cancel debt when it became an independent entity, and as a consequence it now owes the Bristol city utility division nearly $14 million. The auditors' blunt assessment: "These conditions raise substantial doubt about OptiNet's ability to continue as a going concern."

Fiber-Optic Funeral Home

"If you don't have broadband, you can't compete," says Paul Elswick. Elswick's office is located in a repurposed funeral home in an office park in Duffield, Virginia—a setting that would be a little too on-the-nose in a work of fiction. His company, the thematically named Sunset Digital Communications, provides fiber-optic broadband service in the mountain counties of Southwestern Virginia and Eastern Tennessee.

As the BVU Authority's problems mounted, Elswick and Sunset Digital, backed by a Miami-based private equity firm, swooped in and made an unsolicited bid of $50 million for OptiNet in February 2016. No strangers to working within the Appalachian aid-industrial complex, Elswick and his son Ryan have already deployed broadband networks in Southwestern Virginia and Northeastern Tennessee for regional development agencies funded by government grants and loans. In 2010, Sunset applied for federal stimulus funding and received $24.5 million—90 percent grants and 10 percent loans—to construct 279 miles of fiber-optic broadband in Claiborne and Hancock Counties in mountainous Northeastern Tennessee. "Everyone knew that broadband would help the economy in the future, but nobody knew how," says Elswick.

Since there are many local and regional government stakeholders in BVU OptiNet, the process has taken nearly two years to negotiate, but Sunset apparently cleared the final hurdle when the Virginia Coalfield Coalition voted to approve the purchase in August. If the deal holds, Sunset will have bought assets that cost various government agencies $132 million to build for only $50 million—less than 40 cents on the dollar.

This was a smart move for Elswick: Since the public networks have been purchased so cheaply relative to their construction costs, it is highly likely that the new private proprietors will be able to operate them at a profit. In the meantime, any outstanding bond payments will be borne by hapless taxpayers.

'Will Your Bill Go Up?'

A similar story has been playing out in nearby Dickenson County, where the board of supervisors created the Dickenson County Wireless Integrated Network (DCWIN) authority in 2004 as a way to connect businesses, government agencies, and residents to the internet. "Without wireless communications services the county will grow further isolated from the industrial and technical advances" in the rest of the country, the supervisors warned. This, they promised, would "be an outstanding investment for the future of Dickenson County and its citizens."

In 2005, the Board of Supervisors authorized a bond issue of $1.5 million to finance DCWIN's system of 10 high-speed cell towers. The minutes from that public meeting show county resident Gary Harless objecting, arguing that the bond issue would in effect "be mortgaging everyone's property for 15 years." He pointed out that DCWIN at the time had only 150 customers and would need to expand to 1,500 in order to earn the cash to pay off the bonds.

Harless' observations proved prescient. Five years later, the Board of Supervisors dissolved the authority and assumed its debts. A review of DCWIN's budgets since 2009 finds that expenditures always exceeded revenues. In July 2017, the county finally offloaded the wireless network to a local company, Hillcom Inc., for $227,000.

The minutes from a previous Board of Supervisors meeting show that Hillcom founder Brandon Hill had tangled with DCWIN a decade earlier. He'd heard rumors that the authority was trying to put him out of business after he set up high-speed connections to 20 of his neighbors, and he was worried. Looks like he'll have the last laugh.

The Hillcom site minces no words: "Will your bill go up? Well, there was a reason DCWIN was sold. It was not profitable. $39.95 is an extremely low price for internet." The new owner plans to upgrade the service, offering 40–100 megabits per second (mbps) download and 15–60 mbps upload speeds for $100 per month. The company hopes to have 500 customers eventually using its refurbished wireless network. That'll keep it profitable, thanks to the enterprise's ability to grab wireless infrastructure at fire sale rates.

If You Build It

Despite this well-established track record of failure, publicly funded internet infrastructure improvement projects in Appalachia keep getting bigger and more ambitious.

In 2013, Kentucky announced plans to get middle-mile broadband into every one of the state's 102 counties—3,400 miles of fiber by January 2017. The KentuckyWired network was supposed to be finished in the Appalachian counties by April 2016. After $30 million from the state budget, $23.5 million in federal ARC grants, and $232 million in bonds, all the project has to show for itself are 129 miles of not-yet-lighted fiber. In July, Kentucky Communications Network Authority director Phillip Brown flatly declined to set a firm completion date for the entire network.

What's worse, the network threatens to push out private development. Before it began, the state had an agreement with AT&T to bring broadband to Kentucky's 173 public school districts. Democratic Gov. Steve Beshear promised to break that deal to guarantee anchor clients for the network and make the math work to put KentuckyWired in the black—which would also mean the state's taxpayers would foot most of the bill for paying off the bonds. After AT&T threatened to defend its contract in court, the matter was quietly dropped.

This is common. A 2016 analysis from the State Government Leadership Foundation notes that such subsidized broadband networks first remove a major anchor tenant (the government) from private networks, thereby weakening the economic case for private investment. Second, the subsidized networks seek to capture market share from already established private-sector providers. And third, the mere threat of government broadband tends to reduce private-sector investment. Thus, government-subsidized broadband likely impedes rather than speeds up the delivery of broadband service to customers in relatively remote areas.

In 2013, Kentucky announced plans to install 3,400 miles of fiber-optic cable by January 2017. Almost $300 million later, all the project has to show for itself are 129 miles of not-yet-lighted fiber.

"Our big problem with these public-private partnerships is that they never have the private-sector companies carrying most of the risk," says Jim Waters, head of the Lexington-based pro-market Bluegrass Institute. He's right. Macquarie Capital—which holds the contract to build, maintain, and operate the system over a 30-year period—and other private partners are being reimbursed through a fixed set of availability payments over the life of the agreement, regardless of any revenues earned. Kentucky's taxpayers bear the entire risk of revenue shortfalls with respect to the network.

Alarmed by the delays and escalating costs, Republican Kentucky state Sen. Chris McDaniel said at a July hearing, "I want a shutdown plan, with financial costs to shut it down, stop work. What's it going to cost us to get out of this?" McDaniel is right to be concerned. The private partner in a very similar project in Massachusetts, MassBroadband123, filed for bankruptcy earlier this year.

Waters agrees that the state should cut its losses now. "The way things are going, it might take $700 million, $800 million, or even $1 billion to complete the project," he says. Waters also makes the salient point that progress and advancement could make this decadeslong enterprise obsolete. "What about technological change?" he asks. "How do we know that this is the type of infrastructure we will need in 30 years?"

Even as officials were concocting KentuckyWired, access to broadband networks was steadily expanding throughout the state, rising from 85 percent in 2014 to nearly 94 percent in 2016. Waters argues that the bigger problem in Appalachia is not lack of access but the failure to adopt broadband when it's available. This notion is backed up by a 2014 study in The Annals of Regional Science by the Oklahoma State University economist Brian Whitacre and his colleagues. They found that increases in broadband adoption between 2008 and 2011 in non-metro counties did bring increases in income and the creation of new businesses. But "simply obtaining increases in broadband availability (not adoption) has no statistical impact on either jobs or income." If you build it and they don't come, there's little benefit.

Instead of spending hundreds of millions on KentuckyWired, Waters argues, a public information campaign explaining how broadband services can help businesses in Appalachia would be more effective at boosting employment and economic growth.

Coal Miners to Coders

If it's wishful to think you can spark growth with a policy of "if you build it, they will come," it sounds even more fanciful to form a strategy around "if you build it, they will stay." Yet the government has embraced exactly that idea.

The feds think subsidized high-speed internet connections could support newly trained digital workers. In 2015, the Eastern Kentucky Concentrated Employment Program Inc. (EKCEP)—which was still pushing training for coal jobs as recently as 2006—began dispensing federal grants to train mountain folk in the art of computer coding.

The Corridor G highway leading into Pikeville is impeccable. Thanks to the presence of a university and a regional medical center, its downtown, unlike that of many other fading Eastern Kentucky communities, remains relatively vibrant. The electronic sign outside the courthouse proudly declares that Pike County is "America's Energy Capital." In 2016, Fortune listed Bit Source, which is headquartered there, as one of "7 World-Changing Companies to Watch," and in 2017 Fast Company declared its president "one of the most creative people in business."

The outfit is the brainchild of local entrepreneurs Charles "Rusty" Justice and M. Lynn Parrish, who developed the idea in 2014 after a fact-finding trip to a computer-coding incubator in Lexington. Fueled by $150,000 in National Emergency Grant funds from the U.S. Department of Labor, Bit Source selected 10 former coal industry workers out of 900 applicants to be interns. Ranging in age from 33 to 48, they were paid $15 per hour during a 22-week crash course in HTML, CSS, Javascript, and Drupal. All 10 of the selected applicants made it through the training—funded by another $166,000 federal grant—and are still working for the company. Bit Source's software developers now earn from $21 to $23 per hour.

James Johnson, 47, grew up about 8 miles outside of Pikeville. He worked for years selling heavy equipment to coal mining companies for Brandeis Machinery; as the mines shut down, Brandeis downsized and Johnson lost his job. When I meet him at Bit Source's headquarters in a refurbished Coca-Cola bottling plant, I ask why he didn't leave to seek employment elsewhere. "My wife has a good job at the local hospital and my two sons were in school," he replies.

He applied for a lot of jobs at lower wages than he had been earning, but he couldn't get hired. Then Johnson heard about Bit Source and dutifully put in an application without much hope. "By that time, I was so heartbroken and filled with a sense of failure that I didn't think that there was much of a chance that I would be accepted," he recalls. The training was intensive but he found that he could handle it. Did he like coding? "My old job was very routine, very comfort zone." He smiles. "This job is a lot more exciting. You never know what new thing you've got to learn. You sit at your computer and make things out of nothing."

Johnson is convinced that his fellow Appalachians can compete with coders in India, Europe, and South America. "We just need a fast-flowing internet," he says. "We are hoping real hard for the KentuckyWired fiber."

Bit Source Creative Director Payton May, a 29-year-old native of the area, spent two years in architecture graduate school at the University of Virginia studying urban and environmental design. "At 18, I never thought I would be back here," he says. "I now see the value that the area really has. It's home and it's family." One interesting tidbit from May: He says the company's connection to the internet has 15–50 mbps download and 15 mbps upload speeds, well within the parameters of the formal definition of broadband.

The next day, I drove up another congestion-free highway to Paintsville, Kentucky, to talk with several people in a computer training program at the downtown campus of Big Sandy Community and Technical College. The program was being run by Interapt, a Louisville-based software development company that specializes in mobile applications and wearables.

Unlike Pikeville, Paintsville had clearly seen much better days. The main street was mostly deserted and lined with empty storefronts, although a Pokémon Go charging station was attached to a lamppost downtown. Some of the yards sported "Friends of Coal" signs urging people to "Support Kentucky Jobs!"

Interapt's TechHire Eastern Kentucky (TEKY) program involves 16 weeks of intensive training followed by a 16-week apprenticeship at the company. The TEKY program was funded with $2.75 million in grants from ARC, the U.S. Department of Commerce, and the U.S. Department of Labor. Fifty participants were selected from a pool of 850 applicants. None of the Interapt coding trainees had previously been coal miners.

The participants were paid $10 an hour during the training period. "You can't expect people to learn something hard if they are worried about how to feed their families," says Interapt founder and CEO Ankur Gopal. If all 50 completed the program, that would have amounted to $320,000. According to Gopal, between eight and 15 of the company's engineers and designers were typically on site at TEKY. Those staffers charged less per hour than they would for regular client services. Only 33 of the initial 50 students made it through to the apprenticeship phase.

Alex Hughes, 43, grew up in nearby Prestonsburg. He worked for 15 years as a self-employed videographer, often for local law firms. As with much else, the collapse of the coal industry caused that source of work to dry up. He stayed in Eastern Kentucky because "that's where my family is."

Melissa Anderson, 40, grew up in Vergie, near Pikeville. She had worked in administrative positions at a local law firm and then at Big Sandy Community College. Budget cuts at the school resulted in her being laid off in January 2016. She and her fiancé went to Florida for two months looking for jobs, but came back when he could not find steady construction work. She found the TEKY program a "little strenuous" and didn't think she'd make it through to the apprenticeship program. So, taking her business background into account, managers at Interapt offered her a position starting in May as a marketing analyst.

As with the highway construction project before it, the internet infrastructure push has not created a detectable boom. Population in the counties covered by various government-subsidized broadband networks continues to fall.

Lucas Lell, in his early 20s, was the youngest TEKY graduate in the group. Reared in the town of Stopover, he was warned as a boy not to go to work in the coal fields. "You'd be broken by your 40s," his parents told him. Lell had just finished his associate's degree in science at Big Sandy. He was thinking of attending Morehead State, just an hour and a half from Paintsville, when he heard about the Interapt TEKY program. "I've always had a passion for computers," he says. At the end of the apprenticeship program, Interapt offered him a full-time job as a quality assurance analyst. "I do want to stay around here," says Lell. "I'm already far away from my true home, Stopover."

Ultimately, Interapt hired 15 of the TEKY program participants, most of whom work remotely from locations in Eastern Kentucky. Their salaries range from $37,000 to $42,000 a year. Some other participants found tech jobs in the area, but many are still searching. One way to look at the TEKY computer coding program is that it subsidized the training of Interapt's new employees at the rate of $180,000 per hire. Bit Source managed to train folks for considerably less: about $31,000 per employee.

By comparison, the nonprofit Eleven Fifty Academy across the Ohio River in Indiana offers a highly regarded 12-week coding boot camp for $13,500. Students at Big Sandy Community and Technical College can take a year's worth of computer programming classes toward an associate's degree for under $15,000, including tuition, books, room, and board. In July, Gopal suggested that the company would relaunch its TEKY program this fall, but EKCEP has announced that it will not use Interapt in its job training programs in the future.

A Future for the Holler?

I loved meeting the folks in Southwest Virginia and Eastern Kentucky, and I was impressed by Sunset CEO Paul Elswick's business savvy and determination to provide new opportunities to people who live in the region he loves. The drive, enthusiasm, and optimism of the newly minted coders at Bit Source and Interapt was likewise invigorating. Nevertheless, it is hard to see the seeds that are supposed to someday sprout and grow into a nascent Silicon Holler.

It's difficult to tell how many employers, if any, have decided to relocate to Southwestern Virginia due to better access to high speed data networks. As with the highway construction project before it, the internet infrastructure push has not created a detectable boom. Population in the counties covered by various government-subsidized broadband networks continues to fall, dropping from 334,000 in 2000 to 324,000 now. Between 1980 and 2000, by contrast—without any high-speed internet to speak of and with the highways uncompleted—the area's population dropped by a smaller amount, from 336,000 to 334,000.

For more than 50 years, the feds have poured billions in job training and infrastructure funds into central Appalachia with the goal of spurring economic growth and reducing endemic poverty. There is very little to show for all that effort.

In September, I contacted Dickenson County resident Gary Harless, the brave citizen who spoke up at that Board of Supervisors meeting 12 years ago to warn that poorly conceived infrastructure investment would end up "mortgaging everyone's property." I asked him what he thought now. "Looking back, I just feel sorry for the county," Harless told me. "I don't feel smart; I feel like it was just basic economics. Government has never been good at management."