This ADBI Working Paper explores a novel financing framework for large-scale infrastructure, using the High-Speed Rail (HSR) in Taipei,China as a primary case study.
TL;DR
Infrastructure projects often fail to attract private investment because direct revenues (like ticket sales) are too low to cover massive upfront costs. This paper proposes capturing "spillover tax revenues"—the indirect tax gains from increased economic activity around a project—and sharing them with investors through floating bonds and land trust schemes to significantly boost project bankability.
What Problem Does This Paper Address?
Developing Asia faces a staggering $1.7 trillion annual infrastructure investment gap. Traditional funding sources—government budgets and multilateral bank loans—are stretched thin by high national debt and the economic fallout of the COVID-19 pandemic.
The core challenge is the "low-return trap." Essential infrastructure like railways or water systems must keep user charges low to remain accessible. However, these low fees result in a low Internal Rate of Return (IRR), making projects unattractive to "patient" private investors like pension funds and insurance companies. Furthermore, land acquisition is often a bureaucratic and political nightmare that leads to massive cost overruns and delays.
Methods
The authors propose a three-pronged financial innovation:
- Spillover Tax Identification: Using a Difference-in-Differences (DID) econometric analysis, the researchers isolate the increase in regional tax revenue (property, business, and income taxes) specifically attributable to the infrastructure project, compared to a control group without the project.
- Floating Infrastructure Bonds: These bonds pay a base interest rate (equivalent to government bonds) during the non-revenue-generating construction phase. Once the project is operational, the interest rate "floats" upward, fueled by a portion of the captured spillover tax revenues.
- Land Trust Scheme: Instead of the government or developer buying land outright (a massive upfront cost), landowners retain ownership and lease the "right of use" to the project. In return, they receive dividends from the project’s success, reducing initial capital requirements and local resistance.
Datasets
The study utilizes:
- Regional Tax Data: Specific to the HSR corridor in Taipei,China, used to calculate the DID spillover effects.
- HSR Project Financials: Actual cost and revenue data for the Taipei,China HSR (e.g., NT$1,054 billion total cost).
- Japanese Macroeconomic Estimates: Historical data from Japan (1956–2010) used to benchmark the ratio of spillover effects to direct effects (finding that 66–68% of total economic impact is typically indirect).
Key Insights and Ideas
- The 50/50 Revenue Split: The paper suggests a "half-and-half" sharing model where 50% of the incremental spillover tax revenue is returned to the project operators/investors, while the government retains the other 50%.
- Subsidizing the "Quiet" Years: By paying bond interest during the construction phase (T0 to T1), the government provides a safety net that allows private investors to enter projects that would otherwise be too risky.
- The "Virtuous Circle": Improved infrastructure raises the marginal productivity of capital and labor in the region, which increases the tax base, which in turn provides more revenue to pay off the infrastructure—a self-reinforcing loop.
- Land as an Asset, Not a Cost: The land trust model transforms landowners from "obstacles to be cleared" into "stakeholders with an interest in the project's success."
Key Results
The case study of the Taipei,China HSR demonstrates a clear financial uplift when moving from traditional to proposed schemes:
- Traditional Investment (Land Purchase): IRR of 5.1%.
- With Land Trust Scheme: IRR increases to 5.4%.
- With Land Trust + 50% Spillover Sharing: IRR jumps to 7.2%.
- Public Sector Win: Even after the government "subsidizes" the interest during construction (costing NT$409.9 billion), its 50% share of the spillover tax (NT$502.7 billion) results in a net surplus of NT$92.9 billion.
Limitations and Future Work
- Econometric Complexity: While DID is a robust tool, the authors acknowledge that applying sophisticated econometric models to every infrastructure project in real-time is difficult. Simpler "rules of thumb" for tax sharing may be needed for broader adoption.
- Legal Prerequisites: The land trust model requires a mature legal and "Trust Business" framework, which many developing nations currently lack.
- Fixed Tax Rates: The model assumes tax rates remain constant; significant changes in fiscal policy could disrupt the projected returns for the floating bonds.
- Future Direction: The authors suggest exploring how secondary activities, such as Transit-Oriented Development (TOD) and real estate, can further diversify and stabilize revenue streams for infrastructure investors.
Summary updated