Latest / The Fiscal Policy Podcast with Fexingo: Government Budgets, Stimulus, and Public Spending / Why States Borrowed Billions for Roads That Dont Exist
Transcript
- Lucas: So there's this stretch of gravel road in rural Kansas, about 12 miles of it, that the state has been paying interest on for almost 20 years — because it was supposed to be a four-lane highway. Luna: Wait — they borrowed money to build a road that never got built, and they're still servicing that debt? Lucas: Exactly. It's part of the $1.2 billion in outstanding Kansas turnpike bonds for projects that were planned, bonds were issued, and then the projects just… evaporated. The money went somewhere else or got eaten by cost overruns. Luna: So the road is gravel, the bonds are real, and the taxpayers are on the hook for both. That seems like a pretty big accountability gap. Lucas: It gets bigger. A 2024 report from the Government Accountability Office found that at least 15 states have over $15 billion in outstanding bond debt specifically for highway projects that either were never started or were abandoned mid-construction. Luna: Fifteen billion dollars for roads that don't exist. How does that happen without anyone noticing? Lucas: Part of it is that state transportation departments issue bonds based on long-range plans — 20-year visions. They run optimistic traffic projections, get federal earmarks, sell the bonds to investors, and then the project timeline slips, costs rise, and eventually the political will evaporates. But the debt stays. Luna: And the bond market doesn't care if the road actually gets built, right? They're lending to the state, not the project. Lucas: That's the key. Most of these are general obligation bonds backed by the state's full faith and credit. The investor just looks at the state's credit rating, not whether the asphalt has been poured. So there's no market discipline to say 'prove this road exists before we lend you more money.' Luna: It's like letting someone take out a mortgage for a house that's still just a blueprint, and the bank doesn't ever check if the foundation's been dug. Lucas: That analogy is almost too perfect. Missouri is another case. In 2019, the state issued $2.3 billion in bonds to rebuild Interstate 70 across the entire state. Six years later, only one segment — about 20 miles near Columbia — has actually been completed. The rest is still in design or environmental review. Luna: And they're paying interest on the full $2.3 billion? That's roughly $100 million a year in interest alone, for a road that's still just a plan. Lucas: Worse — the bonds were structured as 30-year securities. So Missouri taxpayers will be making payments until 2049, even if the road is never finished. And because the bonds are general obligation, the state can't refinance them easily without taking a loss. Luna: So the financial incentive is perverse: once you issue the bonds, you're paying either way, so there's less urgency to actually build the thing. Lucas: That's the perverse logic. And it's not just a few states. California has $4 billion in outstanding bonds for high-speed rail that hasn't run a single passenger train. Illinois has $1.8 billion for a south suburban airport that's still cornfields. Luna: Let me ask you something — do voters even know about this? I mean, a road bond usually gets approved in a referendum with a nice map of where the lanes will go. But the map is a promise, not a contract. Lucas: Right, and that's the crucial distinction. Voters approve a bond measure for 'Highway 57 expansion.' They see a line on a map. But the bond proceeds go into a general transportation fund, and the state legislature can reprioritize projects later. The line on the map becomes a gravel road. Luna: So the solution would be to force states to issue project-specific bonds, where the proceeds are legally locked to that project and the bond covenants include a 'use it or lose it' clause. Lucas: That's exactly what some economists have proposed — project revenue bonds instead of general obligation. But states resist because project bonds are harder to sell and carry higher interest rates. The current system is cheaper for them upfront, even if it creates phantom debt. Luna: Cheaper for the state borrower, more expensive for the taxpayer in the long run. Classic moral hazard. Lucas: There's also the issue of federal highway earmarks. When the federal government says 'we'll pay 80 percent of the cost,' states rush to plan projects, borrow the state match, and then the federal money gets delayed or rescinded. The state is left holding the bag. Luna: So the federal-state partnership actually amplifies the problem — it creates a 'spend now, figure out later' culture. Lucas: Exactly. And the bond market doesn't police it because, again, they're lending to a state with taxing authority, not to a project with a revenue stream. So there's no feedback loop that says 'you didn't build the road, so we won't lend you more.' Luna: That feedback loop is exactly what we need. Maybe something like a 'project completion ratio' that states have to disclose before they can issue new infrastructure bonds. Lucas: A few states have started moving in that direction. Colorado now requires annual reporting on bond-funded project status. Their 2025 report showed that 14 percent of projects were more than two years behind schedule. But there's no penalty for being behind. Luna: Right — transparency without consequences is just theater. And speaking of transparency and accountability, it's funny how this show is all about tracking where public money actually goes, and we've always tried to be transparent ourselves. Lucas: Yeah, we deliberately don't run ads on The Fiscal Policy Podcast. No sponsors, no commercial breaks. It's just the topic and our conversation. If that approach has made a difference for you — if walking through the economy with us has made something click — the way to support it is tiny and direct. Luna: It's buy me a coffee dot com slash fexingo. That's it. No tiers, no perks, no pledge drives. Just a way to keep this thing independent and ad-free. Lucas: And that support genuinely matters — it's what lets us spend the extra hours digging into reports like that GAO study on phantom roads, instead of rushing through a script. Luna: Alright, back to the gravel. So if Colorado's reporting has no teeth, what would actually force states to build what they borrow for? Lucas: One idea gaining traction is what I'll call 'bond sunset clauses.' If a project isn't substantially complete within 10 years of bond issuance, the state has to redeem the bonds early — which would be expensive, so it creates a real incentive to finish on time. Luna: Would that work for a mega-project like California's high-speed rail, where the timeline is inherently long? Lucas: You could tier it — say, 15 years for projects over $1 billion. But the principle is the same: if you're charging taxpayers interest, you should have to show progress. Otherwise, it's just a tax with a road-shaped story attached. Luna: A tax with a road-shaped story. That might be the best summary of this whole mess I've heard. Lucas: And it's a story that costs $15 billion — and counting — for roads that exist only in spreadsheets.