The Asphalt Ledger – Part 3: The Lock-In Machine — Why the Subsidy Persists and Who It Serves

The City That Removed the Highway and Grew

#

In 1991, San Francisco's Embarcadero Freeway — a double-deck elevated structure that had severed the city's waterfront from its financial district since 1959 — was damaged in the Loma Prieta earthquake. City engineers declared it structurally unsound. In 1991, after a contentious public debate, San Francisco approved its demolition rather than its rebuilding. The freeway had carried approximately 100,000 vehicle trips per day. Transportation planners predicted traffic chaos in the surrounding street network.

The chaos did not materialise. A 2006 study by the University of California Transportation Center found that approximately 30% of the former Embarcadero Freeway's traffic had simply disappeared — not relocated to alternative routes but eliminated, as drivers changed routes or modes in response to the altered road environment. Property values in the blocks adjacent to the former freeway's footprint increased by approximately 300% between 1991 and 2006. The waterfront, which had been commercial and industrial dead space under the freeway's shadow, became one of San Francisco's highest-value mixed-use districts. Auckland, Seoul, Milwaukee, Portland, and Rochester followed with comparable freeway removals — and comparable outcomes.

These cases are not presented here to argue that all freeways should be removed. They are presented because they produce an observable, data-rich natural experiment: what happens when road infrastructure is removed rather than expanded, and what does the outcome reveal about the RSM calculations for roads that were never removed? The answer, consistently, is that the induced demand mechanism operates in reverse as reliably as it operates forward. Traffic that was generated by capacity evaporates when capacity is reduced. And the fiscal obligation that evaporates with it — the 20-year RSM burden the removed road would have accumulated — exceeds the removal cost by factors that the infrastructure finance framework has never been required to compute.

The Political Economy of Locked-In Subsidy

#

Three institutional actors maintain the road subsidy architecture described across this series: the highway construction industry, which captures construction and reconstruction revenue; state DOT bureaucracies, whose budgets and staffing are calibrated to construction management rather than maintenance; and the political incentive structure that rewards infrastructure opening ceremonies and penalises the maintenance spending that prevents failures.

The RSM distribution reveals who benefits from RSM not being calculated. A highway expansion with RSM of 3.5 generates three and a half times its construction cost in future public obligation — maintenance, rehabilitation, and induced-demand-driven early reconstruction. Each dollar of that obligation is a future contract for pavement contractors, engineering firms, and material suppliers. The induced demand term is not merely an economic inefficiency; it is a revenue generator for the construction industry that lobbies for the elasticity assumptions that understate it.

Three Mechanisms That Keep the Subsidy Invisible

#

The Accounting Convention That Separates Construction From Consequence

#

Highway construction projects are authorised and evaluated as capital investments with defined scope, budget, and schedule. Their costs appear as line items in transportation capital budgets. Their lifecycle maintenance cost — the RSM's numerator — appears in future operational budgets, typically recorded separately from capital accounts and managed by different bureaucratic units with different political principals. A state legislator who votes to authorise a $1.6 billion highway expansion is not voting to authorise the $2.2 billion in future maintenance and induced-demand rehabilitation that the RSM analysis would reveal. That obligation will be voted on, in annual maintenance appropriations, by future legislators who inherited it without having created it.

This temporal separation is not accidental. Capital budgets are the arena where transportation policy is made. Maintenance budgets are the arena where transportation policy is serviced. The RSM analysis that would require a capital vote to account for maintenance consequences does not exist as a formal requirement in any U.S. state DOT process or federal project review framework. The FHWA's Benefit-Cost Analysis Guidance for discretionary grant applications requires operationalisation of user cost savings and construction costs. It does not require RSM disclosure.

The comparison with how transit projects are evaluated reveals the asymmetry. Federal transit project grant applications — processed under the Capital Investment Grant programme — are required to submit lifecycle cost analyses that include operating and maintenance cost projections for the full project life, typically 40 years. A light rail project cannot receive federal funding without a prospectus that includes detailed lifecycle maintenance funding plans. A highway expansion project — drawn from the same federal transportation trust fund — is not subject to equivalent lifecycle cost disclosure. Transit is evaluated on lifecycle cost; roads are evaluated on construction cost. This methodological asymmetry systematically advantages road investment in any relative comparison.

The RSM Comparison That Infrastructure Finance Suppresses

#

The Road Subsidy Multiplier framework enables a direct, apples-to-apples comparison between road and transit investment that standard cost-benefit analysis cannot produce, because standard analysis uses different methodological frameworks for each mode.

Under RSM, both modes are measured by the same formula: total 20-year public lifecycle obligation ÷ initial construction cost. Applied to documented projects: Urban highway lane-mile construction costs approximately $8–16 million, with RSM values of 2.1–4.8, producing a 20-year total public obligation of approximately $17–77 million per lane-mile. Metro rapid transit station-to-station construction costs approximately $150–350 million per mile (depending on underground vs. at-grade construction), with RSM values of 0.3–0.8 — transit does not generate induced demand, and its maintenance cost curve is flatter because it does not suffer from heavy-vehicle pavement loading — producing a 20-year total public obligation of approximately $195–630 million per equivalent-lane-mile of capacity.

Transit is more expensive to build per mile. It is substantially less expensive as an RSM-adjusted obligation per passenger-mile of delivered capacity, in any corridor where transit ridership exceeds approximately 5,000 passengers per direction per peak hour — a threshold met in every US major metro corridor where transit has been built. The calculation that the standard framework refuses to make — total lifecycle public obligation per passenger-mile, comparing road and transit on a single standardised basis — is the RSM analysis. It consistently favours transit in dense corridors, and the funding allocation between modes consistently does not reflect that finding.

The Retrofit Cost That Makes Path Dependence Permanent

#

Urban highway infrastructure is, once built, functionally permanent. Not physically — concrete degrades, and the RSM ensures that reconstruction obligations accumulate — but economically and politically. A city that has built its employment geography, its real estate development patterns, and its transit policy around a freeway system has produced a constituency of vehicle-dependent commuters who are exposed to every proposal to reduce road capacity and who will absorb the transition cost of any mode shift. This constituency is not irrational. It is the product of infrastructure investment decisions made over decades that have optimised urban space for the automobile across exactly the geographies where induced demand was highest.

The retrofit cost of reversing car-centric urban design has been documented in a growing body of post-freeway-removal analyses. The Cheonggyecheon Freeway removal in Seoul cost approximately $280 million in 2003 dollars. The subsequent 15-year property value increase in adjacent parcels, the reduction in urban heat island measured at the restored stream corridor, and the traffic redistribution onto transit — ridership on the parallel metro line increased 12% within 18 months — produced a documented return significantly exceeding the removal cost. Rochester's Inner Loop removal (2017, approximately $24 million) generated approximately $200 million in adjacent development investment within five years.

These positive returns are achievable because the RSM-adjusted liability of the removed infrastructure — the future maintenance and rehabilitation that no longer needs to occur — was substantial. Removing a freeway does not merely change urban character. It cancels the future RSM obligation of the structure that was removed. Every year the structure would have operated under an RSM of 3.0 was a year of public obligation that removal converts into freed capital.

The Case for Calculating RSM Before the Shovel Moves

#

The RSM framework does not argue against all road construction. It argues for honest accounting of what road construction costs — not at the ribbon cutting, but across the 20-year horizon that the maintenance, rehabilitation, and induced demand cycles actually operate within.

A highway expansion with RSM of 1.2 — a modest fiscal multiplier, achievable in low-density rural corridors with limited induced demand and efficient pavement maintenance — may represent justifiable public investment. An urban freeway expansion with RSM of 3.5–4.8 — where every construction dollar generates three to five dollars in future public obligation while failing to deliver its projected congestion relief — does not survive honest fiscal analysis. The difference between these two investments is calculable, using data that DOT agencies already collect, in advance of construction. The requirement to calculate it and disclose it does not exist. That absence is not a technical gap. It is a political choice — one that benefits the same industrial interests that lobby the transportation committees who do not require the calculation.

The asphalt ledger has been accumulating for seventy years. The RSM is the accounting instrument that finally lets us read it.