Individual Submission Summary
Share...

Direct link:

When Boundaries Matter: Solar Investment and Spatial Misallocation

Friday, November 6, 8:30 to 10:00am, Property: Boston Marriott Copley Place, Floor: 3rd Floor, Room: Northeastern

Abstract

China's national solar feed-in tariff (FIT), in operation from 2011 to 2019, assigned lower guaranteed electricity prices to regions with better solar resources — a deliberate cross-subsidy intended to equalize private returns across geography. This inverted pricing structure creates a direct tension between the subsidy signal and the social value of generation: each yuan of public support buys the least clean electricity precisely where it is directed most generously. This paper asks whether this spatial architecture distorts the geography of renewable investment, and through which margins.We exploit China's three-tier FIT zone system as a quasi-natural experiment. Zone boundaries are drawn at the county level along provincial lines, creating sharp discontinuities in guaranteed prices between geographically adjacent counties with near-identical natural conditions. Using a matched county-pair panel of 215 border pairs spanning 2013–2022, we combine two-way fixed-effects estimation with a spatial regression discontinuity design to identify the causal effects of the pricing discontinuity.Our data draw on plant-level capacity and generation records covering 19,266 solar installations, official FIT price schedules from five regulatory revisions, and county-level solar utilization hours from China's National Renewable Energy Information Management Platform. The panel spans 3,400+ county-year observations across seven provinces along Zone I/II and Zone II/III boundaries.We find three main results. First, Zone I counties — which receive the lowest guaranteed prices — exhibit approximately 219 more annual solar utilization hours than their Zone II neighbors at the border confirming that the zone classification broadly tracks natural endowment but with substantial imprecision at individual boundaries. Second, despite their superior natural productivity, Zone I counties are significantly less likely to record any new capacity addition: a Heckman selection model finds a log-odds entry penalty of −0.88 (p = 0.022), consistent with marginal investors responding to current price signals rather than long-run site productivity. Third, conditional on entry, the intensive margin — how much capacity is installed — shows no statistically significant difference between zones, suggesting that investors who do enter Zone I are self-selected for longer planning horizons that allow them to internalise the natural productivity advantage. Spatial RDD estimates corroborate an investment discontinuity at zone borders: Zone I counties accumulate more capacity at the Zone I/II boundary (+94 MW, p = 0.039), while Zone III counties accumulate less relative to Zone II at the Zone II/III boundary (−62 MW, p = 0.039). McCrary density tests and placebo zone assignments confirm the validity of the research design.These findings document a measurable spatial misallocation of renewable investment attributable to the FIT's inverted pricing structure. The mechanism operates primarily through investor entry deterrence rather than intensive-margin suppression — a distinction with direct policy implications. Interventions that reduce fixed entry costs in high-resource zones may improve spatial allocation without requiring wholesale reform of the zone-price architecture. More broadly, the results suggest that the spatial design of clean energy subsidies deserves as much attention as their level, and that IRR-equalizing pricing schemes can inadvertently redirect capital away from locations with the highest social return per dollar of subsidy.

Author